Abstract
The descison tree method as a machine learning technique is applied to describe the decision mechanisim in travel behaviour. Four algorithms of decision tree such as ID3, C4.5, fuzzy ID3 and fuzzy C4.5 are introduced to establish the modal choice models. The modelling based on knowledge acquisition from large scale databease would be significant in the description of diversity in decision and the estimation accuracy comparing to conventional approach. Furthermore, it is confirmed that afterward pruning can summarize the information to produce the fuzzy decision tree with variety and high ability of estimation. Finally, the applicability of descion tree based on information gain to the travel behaviour analysis.
